Solving the wicked problem of hospital malnutrition

Hundreds of papers have been published to date outlining the devastating effects of malnutrition on the health and recovery of hospital patients and the difficulties hospital systems face when trying to address the problem. The term ‘wicked’ has recently been used to describe the complexity of the problem of malnutrition and its potential solutions.1 ‘Wicked problems’ were first coined by Rittel and Webber in the urban planning literature to describe problems that are ill defined and poorly understood, that involve stakeholders with differing and sometimes opposing perspectives, and that demand creative solutions, rather than simple or ‘tame’ solutions.2, 3 Wicked problems exist within many aspects of Australian contemporary society, with examples including climate change, indigenous disadvantage and land degradation.4 Thinking of hospital malnutrition as a wicked problem may help dietitians and other health professionals to better understand how to solve this complex problem. The wickedness of malnutrition becomes clear once we consider the characteristics of wicked problems described by Rittel and Webber3 (outlined in Table 1). The first characteristic of wicked problems is no definitive formulation of the wicked problem. Experts have spent decades debating how best to define and diagnose malnutrition. Numerous screening and assessment tools have been developed, validated, refined and re-validated, yet there is still a lack of agreement about the best way to screen for malnutrition and assess nutritional status.5-7 This makes it difficult to definitively report on the magnitude of the problem and conclude on the efficacy of malnutrition interventions. Further complicating the issue of malnutrition assessment and diagnosis are the blurred boundaries between malnutrition (or starvation) and other wasting syndromes such as sarcopenia and cachexia, which are caused by different mechanisms and therefore have different diagnostic criteria and nutrition intervention.8 Guidelines outline parameters for monitoring nutritional status; however, these parameters can be slow to change and confounded by other factors,9 making it difficult to assess the effectiveness of short-term inpatient nutrition interventions. This difficulty in defining and assessing the problem is a key characteristic of a wicked problem. Wicked problems typically exist due to a complex interplay of factors which is poorly understood and difficult to tease out (Characteristic 9, Table 1). Numerous risk factors for malnutrition have been named, including those related to acute and chronic disease, ageing and psychosocial factors. For hospital inpatients, these risk factors rarely present in isolation. Additionally, wicked problems are also often a symptom of another problem (Characteristic 8, Table 1). In the case of hospital malnutrition, poor nutritional status and intake is rarely due to only disease-related, ageing or psychosocial factors, but can be further exacerbated by system factors such as lack of prioritisation of nutrition at ward and organisational level, diffusion of responsibility for nutrition care, and lack of integration of clinical dietetics and foodservice provision.10 These system factors are critical when providing nutritional care in hospital, as this relies on a range of different disciplines contributing to the nutrition care process. For example, the dietitian may prescribe a nutritional supplement or therapeutic diet for a patient, but foodservice officers are responsible for preparing and delivering the correct item, nurses are responsible for assisting patients to consume it, doctors play an important role in reinforcing the importance of following the diet prescription, and so on. This means that nutrition is everyone's responsibility, but often ‘nobody's job’.10 Wicked problems are unique to the context (Characteristic 7, Table 1), with the cause of the problem and response to intervention varying between settings. An example is the difference in baseline practices and response to mealtime programs between hospitals,11, 12 and even between wards within the same organisation.13 The fact that wicked problems vary between context settings means that ‘one size fits all’ solutions are unlikely to be successful, meaning that the task of designing and implementing strategies to address malnutrition is equally as complex as the problem itself. ‘Tame’ or simple solutions are rarely effective when dealing with wicked problems. Therefore, taking a simplistic approach to solving the problem of malnutrition in hospitals is unlikely to result in sustained improvements in nutrition care. There is good research evidence that oral nutrition supplements (ONS) can improve a range of patient outcomes and reduce health-care utilisation,14, 15 making it an attractive ‘one size fits all’ strategy. However, research trials show the provision of ONS is often supported by frequent interactions with dietitians (or researchers) to monitor, and actively encourage consumption of the prescribed supplement. When the provision of ONS in clinical practice is audited, we often find that patients are not provided with the prescribed amounts16-18 or not encouraged to consume them.19 As is common with wicked problems, malnutrition can be a symptom of broader system problems (such as deficits in communication, staff knowledge and/or prioritisation of nutrition),10 and a single-faceted intervention such as supplement provision is unlikely to solve the problem without complementary strategies. Other strategies to address malnutrition may not succeed because they are simple in concept but deceptively difficult to implement. An example of this is Protected Mealtimes where non-urgent interruptions and procedures on the hospital ward are ceased,20 with the assumption that reducing interruptions during mealtimes will improve nutrient intake. Firstly, there is limited evidence that interruptions to meals negatively influence meal intake, with only two studies suggesting a potential link.21, 22 In contrast, other studies have suggested that interruptions to hospital meals may not influence food intake.13, 23 An explanation might be that different types of interruptions affect meal intake differently, and that a short interruption by a staff member can, in some cases, create an opportunity to encourage and assist intake. In one study, Protected Mealtimes unexpectedly resulted in a decreased level of mealtime assistance,23 potentially explained by staff using this time to complete non-meal tasks such as staff lunch breaks. Secondly, there is limited evidence that reducing interruptions during the hospital mealtime is even possible or that it leads to improvement in intake, with only one study out of four that have published on Protected (or ‘assisted’) Mealtimes showing food intake might improve.12 In general, passive strategies are used to support Protected Mealtimes (such as posters, written policies, staff in-services)11, 24 and not surprisingly, minimal change to mealtime care is seen in practice. Where proactive strategies were used as part of an action research approach,12 a reduction in interruptions to the mealtime was not seen, but increases in assistance and intake were observed. These results demonstrate the complexity and unpredictable nature of mealtime interventions as solutions to the problem of malnutrition. If we acknowledge hospital malnutrition to be a wicked problem, we can design and implement strategies to account for this complexity. Isolated approaches such as development and dissemination of policies, standard education packages and ‘toolkit’ approaches are unlikely to successfully resolve the problem without better understanding of the problem and the context in which the problem lies. Implementation science is an emerging area of research defined as ‘the scientific study of methods to promote the systematic uptake of clinical research findings and other evidence-based practices into routine practice’.25 The field of implementation science, also known as knowledge translation, has grown in response to the difficulty in translating research findings into practice, which has been called a ‘slow and haphazard’ process.26 This increased focus on understanding how new health-care knowledge is put into practice, rather than only generating new knowledge about effectiveness of health-care interventions, has led to the development of over 60 published implementation or knowledge translation frameworks.27 These frameworks attempt to make sense of the complex interplay of factors that influence the implementation of innovations into practice, acknowledging that a number and combination of factors influence whether and how innovation is adopted and implemented.28 Commonly used frameworks include Promoting Action on Research Implementation in Health Services (PARIHS,29 recently revised as the integrated-PARIHS30), Theoretical Domains Framework,31 Consolidated Framework for Implementation Research32 and Knowledge to Action.33 Implementing solutions to wicked problems is recognised as being messy, unpredictable and challenging,30 and these frameworks may help to systematically guide researchers and clinicians through the process of implementing complex health-care interventions to address these wicked problems. According to Harvey and Kitson,30 there are four features that are common to most implementation frameworks (outlined in Table 2). The recently revised integrated-PARIHS framework provides guidance to facilitators on how they might operationalise an implementation framework to understand the ‘who’, ‘what’ and ‘where’ in order to successfully implement a complex intervention.30 Table 2 outlines some examples of the issues that need to be identified and understood during the implementation process. These issues can be numerous, complex and time-consuming to identify and understand, and the role of facilitator is to make sense of these issues and enable the recipients to introduce change. This process uses a range of complex skills (see Table 2), which includes team building, conflict management, boundary spanning, networking, negotiating and influencing at the local, organisational and outer context levels. When we think back to common characteristics of wicked problems (outlined in Table 1), we can readily see how these frameworks might help us to understand how to approach the problem and explore potential solutions. Inner (local) context: leadership, culture, past experience of change, mechanisms for embedding change, evaluation and feedback processes Inner (organisational) context: organisational priorities, structure, leadership and management support, systems and processes, culture, history of innovation and change, absorptive capacity Outer context: policy drivers and priorities, incentives, mandates, regulatory frameworks, environmental stability, inter-organisational networks and relationships The Eat Walk Engage (EWE) program is one example where implementation science has facilitated the successful implementation and sustainability of malnutrition interventions into practice. EWE is a collaborative care model for older inpatients, which aims to improve patient outcomes through early mobilisation, adequate nutrition and cognitive stimulation.34 A team of clinician researchers used the PARIHS framework29 to understand and enable recipients to introduce the EWE program on a single acute medical ward. A local multidisciplinary work group (including managers and clinicians) met monthly to develop, implement and reflect on strategies (using consecutive quality improvement cycles consistent with an Action Research approach)35 to improve function, nutrition and cognition of inpatients on the ward. These work group meetings, plus additional service mapping and consultation with stakeholders, allowed the team to gain an in-depth understanding of the recipients and local, organisational and outer context (as outlined in Table 2). Challenges identified by the team included unclear roles and diffusion of responsibility for care tasks, lack of supportive physical environment and resources to support program objectives, workload concerns and risk aversion. A number of enablers were identified including clinical champions across disciplines, previous success with quality improvement projects, existing channels for staff education, and a multidisciplinary team model with well-established communication and governance structures. Specific strategies to improve nutritional intake from a previous ward-based project ‘Encouraging Assisting and Time to Eat’ (EAT)12 were embedded into usual care on the ward as part of EWE. These included increased focus on mealtimes, including: pre-meal preparation, encouragement and assistance during the meal, minimising clinical activity during mealtimes, and provision of higher protein and energy meals, snacks and drinks as standard care. Over an 18-month period, a significant reduction in length of stay was seen on the EWE ward (compared with no change at the organisational level).34 Improvements in feeding assistance previously shown in the EAT trial were maintained throughout the EWE evaluation period.34 Local audit data have demonstrated that the EWE program, plus other hospital-wide foodservice improvements, has sustained high levels of mealtime assistance (100% of patients received assistance required, compared with 58% prior to EAT) and increased daily energy and protein intakes by approximately 900 kJ and 9 g protein per day.35 Varying levels of success have been demonstrated in other studies using implementation frameworks to implement malnutrition programs at the ward, hospital and national levels. At the ward level, a carefully implemented multifaceted nutrition intervention improved nutritional outcomes on an orthopaedic hip fracture unit.36 Action research cycles were facilitated by the ward dietitian (with strong support from the ward geriatrician) to engage ward staff in identifying barriers and enablers within the local context. As a result, a number of strategies were designed and implemented with the aim of improving nutritional intake: promoting nutrition as a medicine, coordinated team approach to nutrition care, improving patient and staff knowledge, and foodservice enhancements (ONS provided as medicine and HPHE mid-meals). After introduction of these strategies, researchers observed that mean energy and protein intakes had doubled, deterioration in nutritional status during admission reduced, and discharge back to the community increased. This was explained by high compliance with ONS, which as prescribed and promoted as medicine by doctors, and supported with positive nutrition messages from the whole team. At the hospital level, the ‘Prevention and Reduction of Weight Loss (PRoWL) in Acute Care’ study used a number of hospital-wide strategies to address malnutrition: nutrition screening, ONS, coloured trays to identify patients needed feeding assistance, staff education and changes to organisational policy.37 Implementation of these strategies was guided by the PARIHS framework29 with dietitian and nurse pairs facilitating the implementation. Evaluation showed successful implementation of the nutrition strategies, with lack of convincing nutritional outcomes possibly explained by difficulties in using weight as an outcome measure in the acute hospital setting. On a larger scale, three different implementation approaches were used to implement perioperative fasting guidelines across 18 National Health Scheme trusts in the UK: (i) guideline dissemination aimed at awareness raising; (ii) guidelines plus web-based resources implemented by local opinion leader; and (iii) guideline implementation using action research cycles with a trained facilitator.38 Over the six-month period, there was no difference in fasting times between the three implementation approaches. Improvements were seen in some individual trusts, where there was a pre-existing organisational commitment to improve fasting times with natural emergence of clinical champions to influence change. However, in areas resistant to change, cultural issues such as inter-professional relationships, communication and culture of innovation were pervasive and difficult to influence in the six-month implementation period. This example nicely demonstrates the strong interplay between context and uptake of an intervention, which can determine the success of implementing evidence into practice. Using an implementation framework which guides the implementer through the process of understanding the what, who and where (Table 2) can inform the facilitation approach required and what implementation strategies may be most useful for the recipients within each unique context. The present study also highlights the significant investment of time often required to understand and successfully influence complex contextual factors at the inner and outer levels.30 The examples discussed here demonstrate the complex interplay between malnutrition interventions (innovation), health professionals (recipients) and the wards and hospitals in which we implement them (context), and how this can influence the success of malnutrition interventions. It is clear from these examples that simple strategies such as introducing a nutrition screening tool into admission paperwork or starting an ONS program are unlikely to achieve significant or sustainable change when it comes to the wicked problem of malnutrition. Calling malnutrition ‘wicked’ does not mean that it is evil or insurmountable; just that it is complex and needs a significant investment of time and energy in creating, implementing and evaluating solutions. Using implementation frameworks may help dietitians and other health professionals to understand the numerous factors influencing the success of complex and system-wide strategies to better manage the problem of hospital malnutrition. The author did not receive funding to write this manuscript. The author has no conflict of interest to declare. The author independently wrote this manuscript. The author would like to thank Dr Lynda Ross, Dr Merrilyn Banks and Dr Alison Mudge for their critical review of this manuscript.